A multilevel approach to understanding healthcare disparities in the nursing home

University Of Florida, Gainesville FL

Investigators

Abstract

DESCRIPTION (provided by applicant): Prior research indicates pervasive racial/ethnic disparities in quality of care among nursing home residents. This study proposes to examine how nursing home residents'race, nursing home segregation and residential segregation influence quality of care in nursing homes. The dissimilarity index will be used to measure nursing home and residential segregation. This index is the most widely use measure of segregation (Iceland, Weinberg, and Steinmetz, 2002). The objective of this study is understand factors that contribute to nursing home segregation, and also understand how various pathways may lead to disparities in quality of care. This study will 1) identify nursing home and market factors that may contribute to nursing home segregation, 2) examine how nursing home residents'race, nursing home segregation, and residential segregation independently influence quality of care, and 3) identify which level (resident, nursing home, or market) accounts for most of the differences observed in quality of care. The goal of this study is to move from documenting racial/ethnic disparities to understanding causal pathways that lead to racial/ethnic disparities. The sample will consist of Medicare and Medicaid certified nursing homes across the United States between the years of 1999 and 2004. Three data sets will be used in this study: On-Line Survey Certification of Automated Records (OSCAR), Medicare Cost Reports, and Area Resource File (ARF). OSCAR will be used to provide facility information about nursing homes. MDS+ will be used to provide information on individual characteristics. ARF will be used to obtain county level variables. The dependent variables will consist of nursing home segregation and quality outcomes (pressure ulcer, bowel decline, and activities of daily living). The primary independent variables of interest are nursing home residents'race, nursing home segregation and residential segregation. A hierarchical linear model will be conducted to examine the role of nursing home and market factors in relation to nursing home segregation. This model will be used because of the continuous nature of the dependent variable (nursing home segregation) and the hierarchical nature of the data. A generalized linear model will be conducted to examine how nursing home residents'race, nursing home segregation, and residential segregation influence quality. A generalized linear model is utilized due to the dichotomous nature of the quality variables. The intraclass coefficient (ICC) will be used to examine the variance among the three levels (resident, nursing home, and market). With an aging population that is becoming more diverse over time it is essential to understand factors that contribute to nursing home racial/ethnic disparities. While initiatives, such as Healthy People 2010, have been implemented to eliminate disparities, literature continues to provide ample evidence that racial/ethnic disparities are widespread throughout the healthcare setting. The goal of this study is to move from documenting racial/ethnic disparities to understanding pathways that may lead to racial/ethnic disparities.